Aspose.OCR is not recognizing?g the text from the supplied recognition? block

Aspose.OCR Product Family
1

Aspose.OCR is not recognizing​g the text from the supplied recognition​ block and is 99% incorrect. Can you please investigate.

This message was posted using Email2Forum by Mudassir Fayyaz.
2

Hi Chuck,

Thank you for considering Aspose products, and welcome to Aspose.OCR support forum.

Sure we will investigate the matter on our end as soon as you provide the sample image along with your source code snippet. Please also mention the platform (.NET/Java) you are currently using.

Looking forward to your response.

3

Hi,

Thanks for your swift reply. Attached please find the OCR source code and an actual size screen capture of the portion of the image in question. The image is initially captured as one page of a pdf file using Aspose PDF. The entire image cannot be posted due to confidentiality concerns.

We are using OCR .net 4.5 version 1.8.0 and we are a licensed owner of Aspose Total.

Also, in the source you will see a few line commented out for using a memorystream instead of a filestream, which would be our preferred method of handling these files, but I could not get it to work. At OCR.Process() I would get an exception that the stream is closed. Not sure what I missed. Maybe you can see and let me know that as well :slight_smile:

Thanks for your swift response!

Chuck
4

Hi Chuck,

Thank you for providing the sample image.

We are able to replicate the issue of incorrect results returned by the latest version of Aspose.OCR for .NET 1.8.0 while processing your provided sample. The image contains text “8000003507 09-Jan-14” whereas Aspose.OCR for .NET reads it as “86oooo3867 09-Ja<-14”, that is incorrect.

A ticket (OCR-33687) has been logged in our bug tracking system to investigate the problem cause, and to provide a fix at earliest. Please spare us little time to properly analyze the issue on our end. In the meanwhile, we will keep you posted with updates in this regard.

Regarding the issue of “Closed Stream”, the said issue is due to the scope of MemoryStream object. As you have used the “Using” statement to initialize the MemoryStream therefore its scope is limited to the “Using” block. You can confirm this by placing the “End Using” statement after calling the OcrEngine.Process method. Hopefully this will resolve the problem.

5

Hi,

We are looking for an update on this issue ticket (OCR-33687) . It has been some time since this was submitted. Aspose Total was a very expense investment and it seems like we cannot even use it.

Thank you in advance for your prompt reply.

6

Hi Chuck,

First of all, please accept my sincere apologies for a bit delayed response.

I am afraid, the ticket logged earlier as OCR-33687 isn’t resolved yet. Upon discussing this matter with the development team lead, I have learned that the major cause of incorrect recognized characters from your sample is the OcrEngine it self. We are currently working on new algorithms for OcrEngine that upon implementation will yield better recognition rate. The said algorithms are in implementation phase at the moment, and we hope to deliver the first iteration in the first quarter of April 2014.

We will keep in touch with you regarding the said ticket. In the meanwhile, if you have other images that are failing the OCR process, we could would recommend you to submit such images to our sample pool, so we could test them against the next major release of Aspose.OCR for .NET, and provide you the test results.

7

The issues you have found earlier (filed as OCR-33687) have been fixed in this update.

This message was posted using Notification2Forum from Downloads module by Aspose Notifier.
8

The issues you have found earlier (filed as ) have been fixed in this Aspose.Words for JasperReports 18.3 update.